Multi-task learning for chemical named entity recognition with chemical compound paraphrasing

Abstract

We propose a method to improve named entity recognition (NER) for chemical compounds using multi-task learning by jointly training a chemical NER model and a chemical compound paraphrase model. Our method enables the long short-term memory (LSTM) of the NER model to capture chemical compound paraphrases by sharing the parameters of the LSTM and character embeddings between the two models. The experimental results on the BioCreative IV's CHEMDNER task show that our method improves chemical NER and achieves state-of-the-art performance (+1.43 F-score).